GlueX Offline Meeting, April 1, 2015

From GlueXWiki
Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, April 1, 2015
1:30 pm EDT
JLab: CEBAF Center F326/327

Agenda

  1. Announcements
    1. global search and replace on wiki: halldweb1 to halldweb (Mark)
    2. CCDB moving to GitHub
    3. Red Hat Enterprise 7 added to nightly builds (Mark)
    4. b1pi and single-track tests (Mark)
    5. can GlueX use idle cores at JLab? (from Sandy Philpott)
  2. Review of minutes from March 18 (all)
  3. Offline Monitoring Report (Kei) Launch Analysis
  4. Action Item Review

Communication Information

Remote Connection

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2015 on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2015/ .

Minutes

Present:

  • CMU: Curtis Meyer
  • FSU: Aristeidis Tsaris
  • JLab: Mark Ito (chair), David Lawrence, Paul Mattione, Nathan Sparks, Mike Staib, Justin Stevens, Simon Taylor, Beni Zihlmann
  • MEPhI: Dmitry Romanov
  • NU: Sean Dobbs

Announcements

  1. Mark executed a global search and replace on the wiki:
    • halldweb1 -> halldweb
    • hdops.jlab.org/wiki -> halldweb.jlab.org/hdops/wiki
  2. Version control for CCDB source files is moving from our Subversion repository to GitHub. Brad Sawatzky had been helping Dmitry get it covered under the JLab contract with GitHub.
  3. Mark has added a node running Red Hat Enterprise 7 to the list of those running the nightly builds. It is his desktop machine lorentz.jlab.org. Many more warnings are being generated from the gcc 4.8 compiler.
  4. Mark threatened to enlist help in getting the b1pi and single-track tests to complete successfully, something that has not happened for several months.
  5. Sandy Philpott has asked can GlueX use idle cores at JLab?.
    • Mark reported that thousands of Data Challenge 3 jobs have been submitted.
    • We will will likely need to do a lot of simulation of 5.5 GeV beam.
    • We had expected to be launching reconstruction jobs by now, back when 10 GeV was the plan.
    • Justin pointed out that simulation will be memory limited since it must be run single threaded.


Review of the GlueX Offline Meeting, March 18, 2015

We looked over the minutes.

GlueX Use of C++11 features and GCC Versions

After consultation among Dmitry, David, Mark, and Richard Jones, it was agreed that we would endorse Dmitry using C++11 features for the CCDB, but limited to those implemented in GCC 4.4. Dmitry will do the modifications necessary for the CCDB build system.

Mark had argued that there are now best-practices-like methods for memory management that are facilitated by the new smart pointers.

Whether to use these language features more widely has not been agreed on. We would need support from the build system.

Offline Monitoring Report

We looked at Kei Moriya's new Launch Analysis pages. These are now being generated quasi-automatically and will be available for new data when it comes.

  • For each launch there are reports of how the jobs performed in terms of execution time and memory usage among other statistics.
  • There are sections which graphically compare each launch to previous launches in several parameters.
  • The fate of the jobs on each file are cataloged for each launch.

Sean was interested in seeing what difference Mike's timing calibrations make for the next launch. Paul remarked that we could launch whenever those calibrations are ready. Mike reported that he is already remaking REST files using his latest calibrations.

OSG Meeting at Northwestern

Sean reported on the recently held annual Open Science Grid meeting.

  • For the past year the group has been focused in getting ready for handling the data from the next LHC run, but now they are turning to better support for opportunistic running. GlueX was mentioned several times in this context.
  • There is a new focus on tools for small independent researchers to allow them to access the grid efficiently. Part of this problem is distributing data for files sizes on the GB scale.
  • FNAL has several experiments on the scale of GlueX that are interested in using the OSG.
  • OSG user support has been restructured. There is a person dedicated to helping smaller groups, in particular consulting about how to deploy the groups resources and aiding them in climbing the learning curve.
  • There was discussion on the next-generation SRM or SRM replacement. The thinking is to do something HTTP-based.
  • There has been interest in deploying temporary grid sites on the Amazon cloud. Amazon is issuing grants to groups that are interested, especially those that develop tools for others to use. The grants are in the form of cloud computing time.

Topics Deferred

  1. GEANT4 conversion
  2. Compute-efficient electromagnetic background inclusion